List your company

Mathematical Reasoning on GSM8K, MATH, SVAMP, SimulEq, AQuA, SAT, MMLU

70.6Accuracy (Aggregate)DEL
14.02428.71243.458.088May 19, 2026

Evaluation results

MethodLinks
DELBase=Qwen2.5-7B, Zero-shot=true2026.0570.681.75687.65765.478.667.9
DIST2LossBase=Qwen2.5-7B, Zero-shot=true2026.0569.281.7568751.863.876.867.6
NTL-WASBase=Qwen2.5-7B, Zero-shot=true2026.0568.881.554.687.153.559.879.165.9
EMOBase=Qwen2.5-7B, Zero-shot=true2026.0568.581.455.887.650.2637566.2
MLEBase=Qwen2.5-7B, Zero-shot=true2026.0568.381.451.6875163.876.866.3
MixCEBase=Qwen2.5-7B, Zero-shot=true2026.0568.381.655.386.953.561.472.766.5
DELBase=DeepSeek-math-7B-Ins., Zero-shot=true2026.0566.177.651.98352.56174.562.1
NTL-WASBase=DeepSeek-math-7B-Ins., Zero-shot=true2026.0564.97751.781.650.259.473.660.8
EMOBase=DeepSeek-math-7B-Ins., Zero-shot=true2026.0564.877.650.980.754.757.969.562.1
DIST2LossBase=DeepSeek-math-7B-Ins., Zero-shot=true2026.0564.876.651.282.553.758.369.561.7
MixCEBase=DeepSeek-math-7B-Ins., Zero-shot=true2026.0564.677.251.281.254.955.970.161.9
MLEBase=DeepSeek-math-7B-Ins., Zero-shot=true2026.0564.476.851.480.952.759.169.160.9
DART-Math-7B (Prop2Diff)Base=DeepSeek-math, Zero-shot=true2026.0563.187.452.683.873.738.656.848.5
MathScale-7BBase=Mistral, Zero-shot=true2026.056175.637.880.675.744.562.350.6
DELBase=Mistral-7B, Zero-shot=true2026.0556.57640.381.746.953.151.845.5
NTL-WASBase=Mistral-7B, Zero-shot=true2026.0555.675.139.983.544.949.249.546.8
DELBase=Qwen2.5-1.5B, Zero-shot=true2026.0555.45742.67448.148.860.956.1
Skywork-OR1-Math-7BBase=Qwen2.5-Math, Zero-shot=true2026.0555.158.580.349.684.3503555
MixCEBase=Mistral-7B, Zero-shot=true2026.0555.175.639.481.945.949.447.146.6
DIST2LossBase=Mistral-7B, Zero-shot=true2026.0554.674.84081.443.650.847.743.7
MLEBase=Mistral-7B, Zero-shot=true2026.0554.275.139.382.444.248.84544.9
NTL-WASBase=Qwen2.5-1.5B, Zero-shot=true2026.0553.855.143.573.346.143.159.156.3
EMOBase=Qwen2.5-1.5B, Zero-shot=true2026.0553.454.542.673.547.74852.755.1
DIST2LossBase=Qwen2.5-1.5B, Zero-shot=true2026.0553.154.241.973.244.645.755.956.5
MixCEBase=Qwen2.5-1.5B, Zero-shot=true2026.0552.952.842.772.743.247.254.557.3
MLEBase=Qwen2.5-1.5B, Zero-shot=true2026.0552.85442.371.644.245.356.455.5
DELBase=CodeLlama-7B, Zero-shot=true2026.054959.935.772.444.446.943.640.1
WizardMath-V1.1-7BBase=Mistral, Zero-shot=true2026.0548.174.529.375.470.725.333.728
DIST2LossBase=CodeLlama-7B, Zero-shot=true2026.0547.659.43471.244.944.539.639.9
NTL-WASBase=CodeLlama-7B, Zero-shot=true2026.0547.459.73572.344.43942.738.5
MixCEBase=CodeLlama-7B, Zero-shot=true2026.0546.159.934.472.64338.234.540.3
Xwin-Math-V1.1-7BBase=Llama 2, Zero-shot=true2026.0545.574.732.38150.626.427.426
Math-Shepherd-RL-7BBase=Mistral, Zero-shot=true2026.0545.482.622.682.445.324.433.227.1
MLEBase=CodeLlama-7B, Zero-shot=true2026.0545.259.4347240.737.834.138.6
DeepMath-Zero-Math-7BBase=Qwen2.5-Math, Zero-shot=true2026.0544.643.750.838.727.244.953.653.5
MetaMath-7BBase=Mistral, Zero-shot=true2026.0542.272.529.477.534.224.833.223.7
CodeLlama-7BBase=Llama-2, Zero-shot=true2026.0524.325.21349.43.52428.626.9
Llama-2-7BZero-shot=true2026.0520.514.62.534.5530.326.829.8
Vicuna-1.5-7BBase=Llama-2, Zero-shot=true2026.0520.213.14.936.64.927.225.930.4
Galactica-6.7BBase=GAL, Zero-shot=true2026.0516.210.22.225.64.225.617.528